import pandas as pd
import re
import matplotlib.pyplot as plt
data=pd.read_csv('net_results_fashionmnist_full.csv')
def preprocess(data):
result=data.copy()
for i in range(1,16):
if(1<=i<=5):
result=result.rename(columns=lambda x: re.sub('net'+str(i)+' ',str(i)+'W',x))
elif(6<=i<=10):
result=result.rename(columns=lambda x: re.sub('net'+str(i)+' ',str(i-5)+'B',x))
elif(11<=i<=15):
result=result.rename(columns=lambda x: re.sub('net'+str(i)+' ',str(i-10)+'D',x))
return result
data=preprocess(data)
test_acc=data.filter(regex=".*testing accuracy")
train_test_acc=data.filter(regex=".*accuracy")
train_acc=data.filter(regex=".*training accuracy")
train_loss=data.filter(regex=".*loss")
def get_model_wise(df,title,figsize=(15,15)):
for i in range(1,6):
df.filter(regex=str(i)+".*").plot(figsize=figsize,title=title+str(i))
def get_category_wise(df,title,figsize=(15,15)):
for i in ['Wide','Base','Deep']:
df.filter(regex="\d?\d"+str(i[0])+".*").plot(figsize=figsize,title=title+str(i))
get_model_wise(test_acc,'test accuracies for model ')
get_model_wise(train_test_acc,'train VS test accuracies for model ')
get_category_wise(test_acc,'Test accuracies for ')
def get_all_graphs_from_csv(csv):
data=pd.read_csv(csv)
data=preprocess(data)
data_name=csv.split('_')[2]
test_acc=data.filter(regex=".*testing accuracy")
train_test_acc=data.filter(regex=".*accuracy")
train_acc=data.filter(regex=".*training accuracy")
train_loss=data.filter(regex=".*loss")
get_model_wise(test_acc,str(data_name)+' test accuracies for model ')
get_category_wise(test_acc,str(data_name)+' Test accuracies for ')
get_category_wise(train_acc,str(data_name)+' Train accuracies for ')
get_model_wise(train_test_acc,str(data_name)+' train VS test accuracies for model ')
get_all_graphs_from_csv('net_results_fashionmnist_full.csv')
get_all_graphs_from_csv('net_results_mnist_full.csv')
get_all_graphs_from_csv('net_results_Emnist_full.csv')
get_all_graphs_from_csv('net_results_ciphar10_full.csv')